The present invention relates to system performance tuning, and more specifically, to cognitive-based dynamic tuning.
Software-based systems require time-consuming tuning operations prior to live deployment. The tuning operations set parameters of the system to appropriate values for expected loads to achieve performance, throughput, response time, and capacity that minimize outages, critical situations, and customer dissatisfaction. The countless and ever-increasing software-based systems include autopilot, guidance, and landing systems in aircraft, navigation and self-parking systems in automobiles, online banking applications, and other consumer applications as examples. Some applications (e.g., airplane landing system, online banking application) may have tighter tolerances on the speed and accuracy of operations than other, less critical applications (e.g., offline backup). Because of scheduling pressures, tuning activities may often be reduced or omitted prior to the deployment of many systems.